1.. SPDX-License-Identifier: BSD-3-Clause 2 Copyright 2018 The DPDK contributors 3 4ABI and API Deprecation 5======================= 6 7See the guidelines document for details of the :doc:`ABI policy 8<../contributing/abi_policy>`. API and ABI deprecation notices are to be posted 9here. 10 11Deprecation Notices 12------------------- 13 14* meson: The minimum supported version of meson for configuring and building 15 DPDK will be increased to v0.47.1 (from 0.41) from DPDK 19.05 onwards. For 16 those users with a version earlier than 0.47.1, an updated copy of meson 17 can be got using the ``pip3`` tool (or ``python3 -m pip``) for downloading 18 python packages. 19 20* kvargs: The function ``rte_kvargs_process`` will get a new parameter 21 for returning key match count. It will ease handling of no-match case. 22 23* eal: To be more inclusive in choice of naming, the DPDK project 24 will replace uses of master/slave in the API's and command line arguments. 25 26 References to master/slave in relation to lcore will be renamed 27 to initial/worker. The function ``rte_get_master_lcore()`` 28 will be renamed to ``rte_get_initial_lcore()``. 29 For the 20.11 release, both names will be present and the 30 old function will be marked with the deprecated tag. 31 The old function will be removed in a future version. 32 33 The iterator for worker lcores will also change: 34 ``RTE_LCORE_FOREACH_SLAVE`` will be replaced with 35 ``RTE_LCORE_FOREACH_WORKER``. 36 37 The ``master-lcore`` argument to testpmd will be replaced 38 with ``initial-lcore``. The old ``master-lcore`` argument 39 will produce a runtime notification in 20.11 release, and 40 be removed completely in a future release. 41 42* eal: The terms blacklist and whitelist to describe devices used 43 by DPDK will be replaced in the 20.11 relase. 44 This will apply to command line arguments as well as macros. 45 46 The macro ``RTE_DEV_BLACKLISTED`` will be replaced with ``RTE_DEV_EXCLUDED`` 47 and ``RTE_DEV_WHITELISTED`` will be replaced with ``RTE_DEV_INCLUDED`` 48 ``RTE_BUS_SCAN_BLACKLIST`` and ``RTE_BUS_SCAN_WHITELIST`` will be 49 replaced with ``RTE_BUS_SCAN_EXCLUDED`` and ``RTE_BUS_SCAN_INCLUDED`` 50 respectively. Likewise ``RTE_DEVTYPE_BLACKLISTED_PCI`` and 51 ``RTE_DEVTYPE_WHITELISTED_PCI`` will be replaced with 52 ``RTE_DEVTYPE_EXCLUDED`` and ``RTE_DEVTYPE_INCLUDED``. 53 54 The old macros will be marked as deprecated in 20.11 and any 55 usage will cause a compile warning. They will be removed in 56 a future release. 57 58 The command line arguments to ``rte_eal_init`` will change from 59 ``-b, --pci-blacklist`` to ``-x, --exclude`` and 60 ``-w, --pci-whitelist`` to ``-i, --include``. 61 The old command line arguments will continue to be accepted in 20.11 62 but will cause a runtime warning message. The old arguments will 63 be removed in a future release. 64 65* eal: The function ``rte_eal_remote_launch`` will return new error codes 66 after read or write error on the pipe, instead of calling ``rte_panic``. 67 68* rte_atomicNN_xxx: These APIs do not take memory order parameter. This does 69 not allow for writing optimized code for all the CPU architectures supported 70 in DPDK. DPDK will adopt C11 atomic operations semantics and provide wrappers 71 using C11 atomic built-ins. These wrappers must be used for patches that 72 need to be merged in 20.08 onwards. This change will not introduce any 73 performance degradation. 74 75* rte_smp_*mb: These APIs provide full barrier functionality. However, many 76 use cases do not require full barriers. To support such use cases, DPDK will 77 adopt C11 barrier semantics and provide wrappers using C11 atomic built-ins. 78 These wrappers must be used for patches that need to be merged in 20.08 79 onwards. This change will not introduce any performance degradation. 80 81* igb_uio: In the view of reducing the kernel dependency from the main tree, 82 as a first step, the Technical Board decided to move ``igb_uio`` 83 kernel module to the dpdk-kmods repository in the /linux/igb_uio/ directory 84 in 20.11. 85 Minutes of Technical Board Meeting of `2019-11-06 86 <https://mails.dpdk.org/archives/dev/2019-November/151763.html>`_. 87 88* lib: will fix extending some enum/define breaking the ABI. There are multiple 89 samples in DPDK that enum/define terminated with a ``.*MAX.*`` value which is 90 used by iterators, and arrays holding these values are sized with this 91 ``.*MAX.*`` value. So extending this enum/define increases the ``.*MAX.*`` 92 value which increases the size of the array and depending on how/where the 93 array is used this may break the ABI. 94 ``RTE_ETH_FLOW_MAX`` is one sample of the mentioned case, adding a new flow 95 type will break the ABI because of ``flex_mask[RTE_ETH_FLOW_MAX]`` array 96 usage in following public struct hierarchy: 97 ``rte_eth_fdir_flex_conf -> rte_fdir_conf -> rte_eth_conf (in the middle)``. 98 Need to identify this kind of usages and fix in 20.11, otherwise this blocks 99 us extending existing enum/define. 100 One solution can be using a fixed size array instead of ``.*MAX.*`` value. 101 102* mbuf: Some fields will be converted to dynamic API in DPDK 20.11 103 in order to reserve more space for the dynamic fields, as explained in 104 `this presentation <https://www.youtube.com/watch?v=Ttl6MlhmzWY>`_. 105 The following static fields will be moved as dynamic: 106 107 - ``timestamp`` 108 - ``userdata`` / ``udata64`` 109 - ``seqn`` 110 111 As a consequence, the layout of the ``struct rte_mbuf`` will be re-arranged, 112 avoiding impact on vectorized implementation of the driver datapaths, 113 while evaluating performance gains of a better use of the first cache line. 114 115 116* ethdev: the legacy filter API, including 117 ``rte_eth_dev_filter_supported()``, ``rte_eth_dev_filter_ctrl()`` as well 118 as filter types MACVLAN, ETHERTYPE, FLEXIBLE, SYN, NTUPLE, TUNNEL, FDIR, 119 HASH and L2_TUNNEL, is superseded by the generic flow API (rte_flow) in 120 PMDs that implement the latter. 121 The legacy API will be removed in DPDK 20.11. 122 123* ethdev: The flow director API, including ``rte_eth_conf.fdir_conf`` field, 124 and the related structures (``rte_fdir_*`` and ``rte_eth_fdir_*``), 125 will be removed in DPDK 20.11. 126 127* ethdev: The legacy L2 tunnel filtering API is deprecated as the rest of 128 the legacy filtering API. 129 The functions ``rte_eth_dev_l2_tunnel_eth_type_conf`` and 130 ``rte_eth_dev_l2_tunnel_offload_set`` which were not marked as deprecated, 131 will be removed in DPDK 20.11. 132 133* ethdev: Update API functions returning ``void`` to return ``int`` with 134 negative errno values to indicate various error conditions (e.g. 135 invalid port ID, unsupported operation, failed operation): 136 137 - ``rte_eth_dev_stop`` 138 - ``rte_eth_dev_close`` 139 140* ethdev: New offload flags ``DEV_RX_OFFLOAD_FLOW_MARK`` will be added in 19.11. 141 This will allow application to enable or disable PMDs from updating 142 ``rte_mbuf::hash::fdir``. 143 This scheme will allow PMDs to avoid writes to ``rte_mbuf`` fields on Rx and 144 thereby improve Rx performance if application wishes do so. 145 In 19.11 PMDs will still update the field even when the offload is not 146 enabled. 147 148* ethdev: Add new fields to ``rte_eth_rxconf`` to configure the receiving 149 queues to split ingress packets into multiple segments according to the 150 specified lengths into the buffers allocated from the specified 151 memory pools. The backward compatibility to existing API is preserved. 152 153* ethdev: ``rx_descriptor_done`` dev_ops and ``rte_eth_rx_descriptor_done`` 154 will be removed in 21.11. 155 Existing ``rte_eth_rx_descriptor_status`` and ``rte_eth_tx_descriptor_status`` 156 APIs can be used as replacement. 157 158* ethdev: The port mirroring API can be replaced with a more fine grain flow API. 159 The structs ``rte_eth_mirror_conf``, ``rte_eth_vlan_mirror`` and the functions 160 ``rte_eth_mirror_rule_set``, ``rte_eth_mirror_rule_reset`` will be marked 161 as deprecated in DPDK 20.11, along with the associated macros ``ETH_MIRROR_*``. 162 This API will be fully removed in DPDK 21.11. 163 164* ethdev: The ``struct rte_flow_item_eth`` and ``struct rte_flow_item_vlan`` 165 structs will be modified, to include an additional value, indicating existence 166 or absence of a VLAN header following the current header, as proposed in RFC 167 https://mails.dpdk.org/archives/dev/2020-August/177536.html. 168 169* ethdev: The ``struct rte_flow_item_ipv6`` struct will be modified to include 170 additional values, indicating existence or absence of IPv6 extension headers 171 following the IPv6 header, as proposed in RFC 172 https://mails.dpdk.org/archives/dev/2020-August/177257.html. 173 174* security: The API ``rte_security_session_create`` takes only single mempool 175 for session and session private data. So the application need to create 176 mempool for twice the number of sessions needed and will also lead to 177 wastage of memory as session private data need more memory compared to session. 178 Hence the API will be modified to take two mempool pointers - one for session 179 and one for private data. 180 181* cryptodev: ``RTE_CRYPTO_AEAD_LIST_END`` from ``enum rte_crypto_aead_algorithm``, 182 ``RTE_CRYPTO_CIPHER_LIST_END`` from ``enum rte_crypto_cipher_algorithm`` and 183 ``RTE_CRYPTO_AUTH_LIST_END`` from ``enum rte_crypto_auth_algorithm`` 184 will be removed. 185 186* cryptodev: support for using IV with all sizes is added, J0 still can 187 be used but only when IV length in following structs ``rte_crypto_auth_xform``, 188 ``rte_crypto_aead_xform`` is set to zero. When IV length is greater or equal 189 to one it means it represents IV, when is set to zero it means J0 is used 190 directly, in this case 16 bytes of J0 need to be passed. 191 192* scheduler: The functions ``rte_cryptodev_scheduler_slave_attach``, 193 ``rte_cryptodev_scheduler_slave_detach`` and 194 ``rte_cryptodev_scheduler_slaves_get`` will be replaced in 20.11 by 195 ``rte_cryptodev_scheduler_worker_attach``, 196 ``rte_cryptodev_scheduler_worker_detach`` and 197 ``rte_cryptodev_scheduler_workers_get`` accordingly. 198 199* eventdev: Following structures will be modified to support DLB PMD 200 and future extensions: 201 202 - ``rte_event_dev_info`` 203 - ``rte_event_dev_config`` 204 - ``rte_event_port_conf`` 205 206 Patches containing justification, documentation, and proposed modifications 207 can be found at: 208 209 - https://patches.dpdk.org/patch/71457/ 210 - https://patches.dpdk.org/patch/71456/ 211 212* acl: ``RTE_ACL_CLASSIFY_NUM`` enum value will be removed. 213 This enum value is not used inside DPDK, while it prevents to add new 214 classify algorithms without causing an ABI breakage. 215 216* sched: To allow more traffic classes, flexible mapping of pipe queues to 217 traffic classes, and subport level configuration of pipes and queues 218 changes will be made to macros, data structures and API functions defined 219 in "rte_sched.h". These changes are aligned to improvements suggested in the 220 RFC https://mails.dpdk.org/archives/dev/2018-November/120035.html. 221 222* sched: To allow dynamic configuration of the subport bandwidth profile, 223 changes will be made to data structures ``rte_sched_subport_params``, 224 ``rte_sched_port_params`` and new data structure, API functions will be 225 defined in ``rte_sched.h``. These changes are aligned as suggested in the 226 RFC https://mails.dpdk.org/archives/dev/2020-July/175161.html 227 228* metrics: The function ``rte_metrics_init`` will have a non-void return 229 in order to notify errors instead of calling ``rte_exit``. 230 231* power: ``rte_power_set_env`` function will no longer return 0 on attempt 232 to set new power environment if power environment was already initialized. 233 In this case the function will return -1 unless the environment is unset first 234 (using ``rte_power_unset_env``). Other function usage scenarios will not change. 235 236* dpdk-setup.sh: This old script relies on deprecated stuff, and especially 237 ``make``. Given environments are too much variables for such a simple script, 238 it will be removed in DPDK 20.11. 239 Some useful parts may be converted into specific scripts. 240